Sensitivity of coherent and incoherent localization methods

This paper presents a comparative study of tile sensitivity of coherent and incoherent localization methods to modeling errors. It is based on the use of performance analysis tools: (i) the Cramer Rao lower bound, that establishes a bound on the precision of parametric estimates: (ii) a generalization of the classical ambiguity function that assesses the global performance under incorrect modeling assumptions, e.g. introduction of biases. Results showing the behaviour of the two methods are presented for specific examples.<>